Fast subway
Fast subway (source: Unsplash via Pixabay)

I recently sat down with Malte Ubl, software engineer at Google and tech lead on the Accelerated Mobile Pages (AMP) project, to discuss web performance in general and accelerated mobile pages in particular. Here are some highlights from our conversation.

What is AMP?

AMP is an open source JavaScript library and validator that make it super easy to create web pages with predictable, reliable, and fast performance. The core promise is: If a document passes the validator, it is also very, very likely to be fast; and if it does not pass, there are straightforward instructions for how it can be fixed.

Another often overlooked aspect of AMP is that it provides an “embedding interface” for web pages. This allows for next-generation native and web applications that can embed content from around the web while maintaining monetization control for the publisher.

Why are they necessary?

I think everyone agrees that content consumption on the mobile web just got too slow. When we started talking to publishers about this, I think everybody was in agreement that something needed to be done and it seems to be working.

What's the best way to get started using accelerated mobile pages?

My favorite way to get started is with Besides that, our canonical docs are definitely the best way to get started.

What are some exciting topics you see on the horizon for web performance?

I recently posted about our plans for using AMP for ads on the Web (not just on AMP pages). I think that reconciling ad tech with web performance best practices will have tremendous impact on the user experience for all of us.

Besides that, I’m super excited about everything service workers and improving performance under Lie-Fi scenarios.

You're speaking at the Velocity Conference in New York this September. What presentations are you looking forward to attending while there?

I am excited to see “AMP: Does it really make your site faster?”. Our data is a clear and very resounding “yes. We are seeing 711ms median load time across the entire corpus referred to from Google Search. But it’ll be interesting to hear what others are seeing.

Article image: Fast subway (source: Unsplash via Pixabay).